Peak or centroid — which parameter is better suited for quantifying apparent marker locations in low‐energy sputter depth profiling with reactive primary ion beams?

Abstract

AbstractSilicon samples containing narrow boron markers were depth profiled using normally incident 0.25–1 keV O2+ ion beams in combination with secondary ion mass spectrometry (SIMS). The profiles revealed significant deviations from ideal delta distributions, either as tails or shoulders. All tracer profiles exhibited the well‐known apparent shift towards the surface, which is due to the initial decrease in erosion rate. The peak positions, $\catcode`@=11\def\curvedhat#1{\vbox{\offinterlineskip\ialign{##\crcr\hfil$\scriptscriptstyle\frown$\hfil\crcr\hfil{#1}\hfil\crcr}}}\curvedhat{z}$, showed a much more pronounced non‐linear energy‐dependent shift than the centroids, 〈z〉, up to a factor of 2 stronger. To evaluate the form of measured profiles, a reduced shape parameter is defined as Δ = 〈z〉 − $\catcode`@=11\def\curvedhat#1{\vbox{\offinterlineskip\ialign{##\crcr\hfil$\scriptscriptstyle\frown$\hfil\crcr\hfil{#1}\hfil\crcr}}}\curvedhat{z}$. For ideal deltas, this parameter can be calculated using the analytical resolution function (rf) after Dowsett et al. Depending on the orientation and magnitude of the tracer distortions, the measured values of Δ differed from Δrf by up to 0.8 nm. This difference is attributed to the fact that the ‘weight’ of the distortions gives rise to a corresponding change of 〈z〉. The results imply that (relative) marker locations are determined more safely from $\catcode`@=11\def\curvedhat#1{\vbox{\offinterlineskip\ialign{##\crcr\hfil$\scriptscriptstyle\frown$\hfil\crcr\hfil{#1}\hfil\crcr}}}\curvedhat{z}$ than from 〈z〉. Copyright © 2001 John Wiley & Sons, Ltd.